Background

Focal cortical dysplasia (FCD) is a heterogeneous form of cortical lesions. Literature defines FCD as a malformation of cortical development, cortical dysplasia, cortical dysgenesis, or neuronal migration disorder. Although the term “cortical dysplasia” accounts for many different focal malformations of cortical development, including heterotrophy and polymicrogyria, FCD is commonly used to describe malformations as described by Taylor [1].

FCD consists of the following various types of changes:

  • Cortical architectural anomalies:

Columnar disorganization, laminar disruption (changes in the six-layer tangential structure of the cortex)

  • Cytological abnormalities:

Hypertrophic neuronal cells: other than those at the typical position at the fifth layer.

Immature neurons: round or oval cells with large nuclei that are not present in the mature cortex.

Dysmorphic neuron: abnormal size and morphology of axons and dendrites, in addition to an increased accumulation of neurofilament proteins.

Balloon cells: these cells are pathognomonic for Taylor type dysplasia [1]. They generally have the immunohistochemical properties of neuronal and glial cells, including a poorly defined membrane, single or multiple eccentric nuclei, and are large cells with an eosinophilic cytoplasm [2].

FCD was first differentiated from developmental malformations in ten patients by Taylor et al. in 1971, and since then, its classification has undergone various modifications. The palmini classification, which was modified by Blumcke in 2011, has been used most recently (Table 1).

Table 1 New classification system of focal cortical dysplasia by Blumcke et al. 2011
Full size table

FCD is frequently discovered during neuropathological examinations of pediatric patients undergoing surgery (especially those surgeries due to drug-resistant epilepsy), and it is one of the most common causes of partial epilepsy that can be treated with surgery [3, 4]. Gender does not play a role in the development of FCD. The absence of a known genetic cause for mild cortical malformations suggests that their etiology may be due to an external factor that influences the migration and differentiation stages of neurons during cortical development.

The etiology of FCD is not well understood. While it is responsible for approximately half of the drug-resistant epilepsy cases in children and adults, patients with FCD often respond well to treatment.

The continuous improvement of neuroimaging techniques has led to more accurate evaluations of pathological lesions, which creates a better chance that appropriate patients will be recommended for surgical treatment.

Because there has been a rapid development in the fields of genetics and diagnostics, FCD should be reviewed by these disciplines, which may lead to an increased number of treatment methods for this disease.

MRI findings

Magnetic resonance imaging (MRI) is the most common imaging method used to assess brain pathology in FCD.

MRI findings in FCD depend on both the size and the characteristics of the lesions. For example, FCD type Ia causes mild hemispheric hypoplasia without other visible lesions [5, 6]. In FCD type IIb, there are localized areas with increased cortical thicknesses and signs of blurred funnel-shaped regions, which indicates gray-white matter separation at the bottom of the sulcus (transmantle MR sign). Lesion signal intensity in FCD is age-dependent. In neonates and infants, lesion signal intensity in FCD type IIb is hyperintense in T1A images and mildly hypointense in T2A images. In older patients, lesion signal intensity has a hyperintense wedge-shape extending to the subcortical and deep white matter in FCD T2 and FLAIR images. In some cases, the subcortical linear or curvilinear T2/FLAIR hyperintensity focus extends to the superolateral margin of the lateral ventricle [5, 6].

In FCD type IIb, contrast does not hold in T1-contrasted series. Increased diffusion and reduced fraction anisotropy (FA) are seen in FCD type IIb diffusion-weighted images (DWI). Reduced N-acetylaspartate:creatinine ratio (NAA:Cr) and increased Myoinositol (ml) rates can be monitored via magnetic resonance spectroscopy (MRS). Normal or decreased cerebral blood flow (rCBV) is seen in perfusion MR [7].

FCD type III has been recently identified and is primarily encountered in patients with hippocampal sclerosis. FCD type IIIa is characterized by volume loss in the anterior temporal lobe with abnormal white matter hyperintensity on T2 and FLAIR images, while the cortex appears normal from other views [8].

Voxel-based morphometry and statistical parametric mapping are advanced techniques that can enhance the detection of epileptogenic lesions in patients with negative tissue analysis as determined by standard MR images [9].

The current study aimed to determine the prevalence of previously described MR imaging criteria for FCD, to identify MR imaging findings (with the goal of optimizing the detection of FCD in clinical practice), to measure the mean ADC values, and to compare these values with those of cortical-subcortical lesions to determine whether this parameter will be useful in the differential diagnosis of FCD.

Materials and methods

Patients

In this study, we retrospectively reviewed the cranial MRI images of 63 patients (33 males, 30 females) who were suspected of having FCD based on clinical findings (ictal-video electroencephalography (EEG)) and MRI findings, who were resistant to medical treatment, and who were seen in our clinic between February 2011 and July 2016. The age range of the patients was 1–60 years (the mean age was 18.92 ± 16.13 years in males and 25.41 ± 11.79 years in females).

Patients with cerebellar cortical dysplasia, artifacts, and lesions detected in the vicinity of FCD, or patients with suspected FCD in MRI but who were not clinically diagnosed with FCD were excluded from this study.

This study was approved by the ethics committee of our university.

MRG protocol

Brain MR scans were performed with 1.5-T and 3.0-T MRI devices (Achieva; Philips Medical Systems, Best, the Netherlands) using an eight-channel cranial coil. All examinations were performed according to the epilepsy protocol. The MRI protocol routinely used for epilepsy in our hospital includes T2-weighted FSE (fast spin echo) on the axial and coronal planes, fluid-attenuated inversion recovery (FLAIR), and non-contrast T1-weighted 3D turbo field echo (TFE) sequences (Tables 2 and 3). During the examination, T1-weighted axial, coronal, and sagittal plane images were obtained in 3D TFE sequence using intravenous paramagnetic contrast material (gadodiamide [Omniscan; Amersham Health, Cork, Ireland], gadopentetate dimeglumine [Magnevist; Schering AG, Berlin, Germany], and gadobutrol [Gadovist; Schering AG, Berlin, Germany]) at a dose of 0.1 mmol/kg through the antecubital vein at a rate of 2 ml/s based on determined pathology.

Table 2 Parameters used in 3-Tesla MRI
Full size table
Table 3 Parameters used in 1.5-Tesla MRI
Full size table

The cranial MRI images of all patients who were accepted as having FCD via cranial MRI and clinical data (ictal-video EEG) were examined retrospectively by two neuroradiologists (with 5 and 8 years’ experience).

The following five semiologic findings were analyzed:

  1. 1.

    Cortical thickening (CT): thickening of the normal cortex by 50% in at least two sequences (T1WI and T2WI)

  2. 2.

    Cortical signal intensity (CSI): signal change involving the entire thickness of T1WI, T2WI, and/or FLAIR cortex

  3. 3.

    Blurring in the gray-white matter (GWM) interface (blurring): signal change in at least one sequence and two planes

  4. 4.

    Subcortical white matter hyperintensity (SCH): abnormal signal change in the subcortical white matter according to the normal cortex

  5. 5.

    Transmantle sign (TMS): signal change extending from the subcortical white matter to the ventricle in T1WI and T2WI sequences

The DWI images of the lesions detected in routine MRI were analyzed using the Philips Extended MR Workspace. Since the DWI has poor resolution, there may be some difficulties in lesion imaging. In order to prevent this, the lesion should be identified with both conventional and dynamic MRI and evaluated accordingly. A round or oval ROI was placed manually over the lesion to cover at least 50% of the lesion size using b:0 images, which have better anatomical detail compared to other maps. The mean ADC values in these areas were calculated directly on the ADC maps.

The mean ADC measurements were calculated for each patient using the same ROI areas as follows:

  1. 1.

    On the cortical, subcortical visible hyperintense lesion area

  2. 2.

    From the contralateral symmetric healthy parenchyma

Cerebrospinal fluid (CSF) was not included in the measurements in order to prevent false measurements.

Statistical analysis

All data were analyzed using the SPSS 18.0 package program. The Shapiro-Wilk test was used to determine the normality of lesions as well as the mean ADC and contralateral symmetric healthy ADC values. For paired comparisons, we used the parametric “Student’s t test” for groups with a normal distribution and the “Mann-Whitney U test” for groups with a non-normal distribution. The ADC values of the groups with and without lesions were compared according to the test results. Values of p < 0.05 were accepted as significant.

Findings

The male/female ratio of the 63 patients was 1.1 (male = 33; female = 30). The localization of FCD was frontal lobe in 38 (60%) patients, parietal lobe in nine (14%) patients, temporal lobe in seven (11%) patients, insular cortex in four (6%) patients, parahippocampal gyrus in two (3%) patients, and hemispheric area in three (4.7%) patients.

A total of 25 lesions (39.5%) were located in the right hemisphere (frontal lobe in 17 patients, parietal lobe in two patients, temporal lobe in three patients, insular cortex in one patient, hemisphere in two patients), 32 lesions (50.7%) were located in the left hemisphere (frontal lobe in 16 patients, parietal lobe in seven patients, temporal lobe in four patients, insular cortex in two patients, parahippocampal in two patients, and hemisphere in one patient), and six lesions (9.5%) were located bilaterally (frontal lobe in five patients and insular cortex in one patient) (Table 4).

Table 4 FCD anatomic distribution
Full size table

In the current study, 20 patients (32%) had all five of the cranial MRI findings. We found CT-CSI in 40 patients (63%), blurring-SCH in 50 patients (79%), and blurring SCH-TMS in 29 patients (46%).

CT was limited to the small cortical area in 48 patients (76%), and it was not more than twice the normal cortex thickness.

CSI was present in 41 patients (65%). Forty of these patients (97.5%) also had CT.

Blurring was present in 52 patients (82.5%). Of these patients, 50 (96%) had SCH.

SCH was detected in 52 patients (82.5%). Of these patients, 50 (96%) had blurring.

TMS was present in 29 patients (46%). All of these patients had SCH (100%). All of the FCD cranial MRI findings are summarized in Table 5.

Table 5 Distribution of FCD cranial MRI findings
Full size table

All of the lesions presented with a diffusion increase in DWI. The FCD mean ADC values were 1.087 × 10−3 mm2/s (0.82–1.316 × 10−3 mm2/s), and the symmetrical normal parenchyma mean ADC was 0.758 × 10−3 mm2/s (0.678–0.872 × 10−3 mm2/s). The normalities of the mean ADC values of the lesions and of the contralateral healthy parenchyma were investigated using the Shapiro-Wilk test, which indicated that both groups were normally distributed (p < 0.05). The Student’s t test revealed a significant difference between the mean ADC values of the groups with and without lesions (p = 0.001).

The average ADC (c-ADC) values, which were measured from the cortical and subcortical hyperintense areas and contralateral symmetric healthy areas in 63 patients with FCD, are summarized in Table 6 (Figs. 1, 2, 3, 4 and 5).

Table 6 ADC values obtained from lesional and symmetric healthy parenchyma
Full size table
Fig. 1
figure 1

A 48-year-old male patient with epilepsy. a Increase in cortical signal intensity in axial FLAIR sequence-subcortical signal intensity. b Cortical thickening in axial T1A sequence. c Cortical signal intensity in coronal T2A sequence, blurring in white-gray matter interface. d Hyperintense area in axial dADC sequence, ADC measurement from this area and from a symmetrical area

Full size image
Fig. 2
figure 2

A 35-year-old male patient with epilepsy. a Cortical signal change in axial FLAIR sequence-blurring in GW matter interface (arrow) and transmantle sign (star). b Cortical signal change in coronal FLAIR sequence-blurring in GW matter interface (arrow) and transmantle sign (thick arrow). c Increase in cortical signal intensity-blurring in GW matter interface. d Hyperintense area in axial dADC sequence, ADC measurement from this area and from a symmetrical area

Full size image
Fig. 3
figure 3

A 36-year-old female patient with epilepsy. a Cortical thickening in axial T1A sequence. b Subcortical hyperintense in axial FLAIR sequence-blurring in GW matter interface. c Increase in cortical signal intensity in coronal T2A sequence-blurring in GW matter interface. d Hyperintense area in axial dADC sequence, ADC measurement from this area and from a symmetrical area

Full size image
Fig. 4
figure 4

A 13-year-old male patient with epilepsy. a Subcortical hyperintense in coronal FLAIR sequence. b Cortical thickening in sagittal T1A sequence. c Increase in cortical signal intensity in axial FLAIR sequence. d Hyperintense area in axial dADC sequence, ADC measurement from this area and from a symmetrical area

Full size image
Fig. 5
figure 5

A 21-year-old female patient with epilepsy. a Subcortical hyperintensity in axial FLAIR sequence. b Increase in cortical signal intensity in axial T2A sequence and subcortical blurring. c Cortical thickening in axial T1A sequence. d Hyperintense area in axial dADC sequence, ADC measurement from this area and from a symmetrical area

Full size image

Discussion

MRI is the most commonly used imaging method for evaluating brain pathology in FCD. Cranial MR imaging usually shows abnormalities in FCD type 2, but only shows abnormalities in some FCD type 1 cases [1]. Despite the use of appropriate protocols in specialist centers, MRI cannot be used to diagnose type 2b FCD in 50% of patients, which can lead to delays in treatment [4]. Despite severe and persistent epilepsy, in most series, the average time from the onset of seizures to surgery is about 1 year [5,6,7,8]. Previous studies have identified typical MR imaging features such as cortical anomalies (thickening, increase in T2 signal), subcortical white matter (blurring in gray-white matter interface), increase in subcortical signal, and transmantle sign [10,11,12,13,14,15,16,17,18,19,20,21]. Earlier reports of MR imaging findings were based on a limited number of patients [12, 15]. However, the newer and more extensive series have not provided a comprehensive analysis of all reported MR imaging findings, and the prevalence of each finding has varied considerably between studies (Table 7). In addition, the ratio of FCD-negative MR imaging findings has varied significantly due to differences in imaging protocols and selection bias.

Table 7 Typical MR imaging findings of FCD type 2 that have been reported in the literature and histologically proven
Full size table

In the current study, which investigated the cranial MRI findings of FCD, the systematic qualitative basic MRI findings are as follows: 20 patients (32%) had all five of the cranial MRI findings. There was cortical thickening and an increase in cortical signal intensity in 40 patients (63%), subcortical hyperintensity and blurring in the gray-white matter interface in 50 patients (79%), and blurring in the gray-white matter interface, subcortical hyperintensity, and transmantle sign in 29 patients (46%).

Cortical thickening, subcortical hyperintensity, and blurring in the GW matter interface are considered major findings in terms of the MRI diagnosis of FCD.

The increase in cortical signal intensity is a well-known finding of FCD. In our current study, this was found in 65% of the patients, while other studies have reported the increase in cortical signal intensity was found at rates between 15 and 62% [10, 11, 13, 15, 23]. This finding was more obvious in FLAIR sequences. With the increased use of 3D FLAIR and high-field MR imaging, we believe that the cortical signal intensity will be more easily detected and more reliable in the future [23].

The transmantle sign was positive in 46% of the patients in our current study, while previous studies report this range between 20 and 83% [10, 11, 13, 15, 23]. This wide range suggests that the transmantle sign was either ignored or combined with change in the subcortical signal in some studies. It has been reported that the transmantle sign is a specific feature for cortical development malformations [18]. The transmantle sign is also associated with the presence of hypomyelination and balloon cells in white matter underlying the dysplastic lesion [12, 24]. It may also occur in other developmental abnormalities such as venous or arteriovenous malformations, and is not specific in FCD [25]. However, its relationship with cortical thickening and blurring in GW matter is an important sign for the diagnosis of FCD.

In a study conducted with 71 patients (male/female ratio of 1:3) who were diagnosed with FCD type 2b using histopathological correlation, there was a left/right hemisphere ratio of 1:35, and lobar distribution was as follows: frontal lobe (n = 60), parietal lobe (n = 7), temporal lobe (n = 2), and occipital lobe (n = 2). Moreover, subcortical hyperintensity and blurring in GW matter interface was 90%, transmantle sign was 83%, cortical thickening was 71%, and cortical signal increase was 50%. Subcortical signal changes were detected as major findings [23]. In our current study (male/female ratio of 1:1), the left/right hemisphere ratio was 1:28, and lobar distribution was as follows: frontal lobe (n = 38), parietal lobe (n = 9), and temporal lobe (n = 7). Our current findings were similar to those in the literature in terms of gender, hemispheric distribution, and lobar distribution. Although the subcortical signal changes in our current study were detected at lower rates than those found in the literature, the subcortical signal changes with cortical thickening were detected as major findings. The difference between the rates of subcortical signal changes in our current study and in the literature may be due to the fact that our current study only included patients with positive MRI findings.

The findings of frontal lobe dominance in FCD localization and the rare occurrence of lesions in the temporal lobe in our current study are comparable with those in the literature [14, 15, 22]. Other epileptogenic lesions are located predominantly in the temporal lobe [23]. This increases the suspicion that FCD, one of the most common causes of drug-resistant partial epilepsy, may be located in the frontal lobe. This hypothesis has been supported by cryptogenic partial epilepsy series, in which 40% of resected cortical lesions (especially in the frontal lobe) were type 2 FCD via histology [26,27,28].

ADC values are the quantitative correspondence of diffusion in biological tissues. The movement of water molecules throughout tissue is influenced by the tissue’s cellularity and histological structure. These parameters affect diffusion and perfusion, and thus ADC values [29]. Low ADC values are indicative of restricted diffusion due to hypercellularity [30].

In our current study, there was a significant difference between lesional and contralateral symmetric healthy parenchyma ADC values (p < 0.001). The mean ADC value at lesion level was 1.087 × 10−3 mm2/s (0.82/1.316 × 10−3 mm2/s) and the symmetrical normal parenchyma mean ADC value was 0.758 × 10−3 mm2/s (0.678/0.872 × 10−3 mm2/s). FCD may be confused with low-grade astrocytomas with cortical involvement. The mean ADC values reported in the literature for low-grade astrocytomas are quite variable [31,32,33] (Table 8). Although the mean ADC values obtained in our current study differ from some in the literature, it may not always be reliable to distinguish FCD from other lesions by looking only at mean ADC values. The mean ADC values with conventional MRI findings may contribute to a diagnosis.

Table 8 The glial malignancy mean ADC values reported in the literature ×(10−3 mm2/s)
Full size table

In our current study, which evaluated patients diagnosed with FCD both radiologically and clinically, major limitations include a lack of histopathologic correlation, a lack of inclusion of MRI negative patients, and a lack of exclusion of lesions that may be confused with FCD. Furthermore, the specificity of MR imaging findings could not be assessed due to the lack of a control group. In addition, MR imaging findings relied on conventional visual analysis. For this reason, experienced neuroradiologists were used to diagnose FCD with conventional MRI.

Conclusion

SCH, blurring, and CT were the most common findings in our current study. We believe that further investigation of these findings may lead to an easier diagnosis of patients with suspected FCD. All of the lesions in the current study had a diffusion increase in DWI. The quantitative mean ADC values determined via the differential diagnosis of other cortical-subcortical lesions with a diffusion increase can be used as a reference. Detection of conventional MRI findings of FCD, which is resistant to medical treatment, in addition to findings from EEG and dADC values may facilitate the pre-operative diagnosis of FCD.